Revision history [back]

If you have no clue on how to handle this, I think you should start reading a book on two things

1. Computer Vision
2. Programming openCV for Android

Next a good approach on how to solve this could be to

1. Trying to define lines in the image
2. Use those lines to define planes in the image
3. Use those planes to do estimation of dimensions

However, since this is a 1 camera solution, the first main problem will be to have a single camera calibration for correct mapping of pixel dimensions to real world measurements.

Recent research on this can be found here. However this is not predefined in OpenCV, so it won't be just a straightforward call to a function.

You will also have to consider distortion and such, but I wish you good luck with the project!